U.S. short- and mid-term yields are edging higher, while the long end and Canadian bonds are under pressure.

Interest Rate Weekly · 2026-09-07 to 2026-09-13

One-line summary: The main theme over the past week was clear: global inflation concerns resurfaced, tightening expectations for U.S. and European central banks were repriced, long-end yields continued to surge amid a global bond sell-off, U.S. and Canadian curves steepened in sync, but predictability remains concentrated mainly in the U.S. short and mid curve.

1. Review of Last Week

The U.S. story had the highest information density and was most directly reflected in the curve. The news flow evolved from early-week headlines that September would be unfavorable for both stocks and bonds and that U.S. job growth was pushing up rate-hike bets, to midweek successive repricing around CPI, PPI and Fed decisions. On September 8 there were statements that commodity inflation was running hot but the Fed was still focused on core inflation; on September 10 there were reports that the CPI report heavily weighed on the Fed decision, that PPI showed rising prices between firms, that the Fed was closer to hiking, and that the U.S. economy could withstand two rate hikes. By Thursday and Friday market reactions were no longer limited to short-end policy-path adjustments; the entire U.S. Treasury curve faced selling pressure: bond market gets spooked, yields rose across the board, 10-year near 5%, 30-year to 5.37%, and selloff shows no sign of diminishing. This is fully consistent with the data. Over the past 5 trading days, U.S. 2-year, 5-year, 10-year and 30-year yields rose by 17, 21, 16, and 10bp respectively; over 20 days the rises are 36, 37, 27, and 13bp; over 60 days they reach 49, 57, 48, and 40bp. The one-year z-scores for all tenors are above 2.5, indicating this round of increases is in the high tail of the past-year distribution, with the 5-year and 10-year most pronounced at z-scores of 2.75 and 2.77.

The U.S. curve shape also showed noteworthy changes. US 2s10s is currently +39bp, clearly positively sloped. Given that the 2-year and 10-year have risen by 49bp and 48bp over the past 60 days respectively, the absolute magnitudes are similar, but the near-20-day move has been stronger at the short end; later in the week the long end was resent under selling pressure as it approached the 5% threshold. This suggests the market action in this round is not a single factor of policy-rate lifting or term premium expansion, but two alternating forces: the first half was an upward revision of the Fed path, the latter half was broader inflation and supply pressures pushing on the long end.

The Canada story had less news than the U.S. but was not directionally independent. Relevant intraday news included that the Canadian dollar weakened with the oil price pullback and that U.S. CPI did not further strengthen the dollar, and that Canadian mortgage-backed securities obtained a U.S. disclosure exemption. The former indicates market attention to Canadian assets is not limited to domestic rates but includes oil and FX; the latter is more structural and regulatory with limited information about medium-term rate direction. The valuable signal was the curve behavior: Canadian 2-year, 5-year, 10-year and 30-year yields rose by 20, 21, 14, and 10bp over the past 5 days, by 33, 33, 25, and 19bp over 20 days, and by 55, 59, 53, and 44bp over 60 days. The intensity is very similar to the U.S., and z-scores are even higher, all between the 2.88 level and above, with the 5-year at 3.04, showing Canadian rates are already at more extreme high quantiles. CA 2s10s is +63bp, steeper than the U.S., indicating the Canadian curve is concentrating repricing of growth, inflation and risk premia between the front/mid section and the 10-year point.

The global narrative last week was jointly centered on inflation returning and a global bond sell-off. On September 12 there were headlines that inflation is making a comeback globally, bringing a new round of anti-inflation efforts and asking what is driving the global bond sell-off; in Europe, on September 11 the ECB hiked rates because inflation rose, and on September 13 the ECB President said the current inflation shock will last longer than expected; in the U.K., on September 8 there were reports that following gilt selling government borrowing costs reached their highest since 1998 and that Bailey warned a disruption from an Iran war could disturb the U.K. economy and bring inflation risk. These items echo the synchronous rises in U.S. and Canadian curves: global markets are reaccepting a more uncomfortable reality that secondary inflation shocks from energy and commodities remain, the core inflation path has not given long-bond holders much comfort, and term premia in major markets are being pushed up.

Information on credit spreads is limited. U.S. HY OAS tightened 1bp on a 20-day basis and IG OAS widened 1bp, both small changes. At least from this dataset, last week’s rate increases were mainly driven by the risk-free rate side rather than a simultaneous notable deterioration in credit risk. In other words, this looks more like duration repricing rather than a credit-event-driven risk spillover.

2. Views on Future Rates

First, method boundaries. This report gives directional judgments only on signals labeled as proven; for tenors tagged as ≈random walk, direction is unpredictable and no judgment is offered, only static carry and curve position are discussed. The medium holding period is about 3 weeks, which matters: carry can give a tailwind, but it only holds when yields are unchanged or volatility is controllable; if yields continue to rise, price losses can fully overwhelm carry, so carry alone should not be treated as a buy signal.

Start with the U.S. short and mid curve, where directional judgments are possible. Momentum signals for US 2Y, 5Y and 10Y are all marked as [proven], and current direction for each is up. Combining the current curve snapshot, the 2-year yield is 4.56% with 5-day +17bp, 20-day +36bp, 60-day +49bp; the 5-year is 4.75% with 5-day +21bp, 20-day +37bp, 60-day +57bp; the 10-year is 4.95% with 5-day +16bp, 20-day +27bp, 60-day +48bp. The evidence for persistent upward moves in the short and mid curve is relatively complete: first, the news flow intensively discussed higher Fed tightening probabilities and higher-for-longer; second, curve momentum is positive across 5-, 20- and 60-day horizons; third, z-scores are significantly elevated but no data-validated reversal signal has appeared. For the next roughly 3 weeks, a continued bias toward higher yields for U.S. 2-year, 5-year and 10-year is the baseline judgment supported by current data.

Among the three, the 5-year momentum is strongest. Its 60-day rise of 57bp exceeds that of the 2-year and 10-year, and its z-score of 2.75 is high. From a trading perspective, the 5-year sits at the intersection of policy-path and medium-term inflation pricing; over the past week the market revised up both, so it is most sensitive to news shocks. The 2-year is also judged up but is more directly driven by Fed expectations; the 10-year faces the combined influence of policy rates, fiscal supply, term premium and the global long-bond sell-off, so although it can be judged biased higher, its volatility sources are more diverse and its path may be more tortuous.

The U.S. 30-year cannot be given a directional judgment. Data explicitly marks US_30Y≈random walk, and its hit rate has not been materially better than the base rate. Last week the 30-year rose to 5.37% and headlines noted 30-year to 5.37% and long-end yields spiking, but these only describe what has occurred and cannot be extended to predict continuation over the next 3 weeks. For the 30-year, direction is unpredictable and no judgment is offered.

For the long end only carry can be discussed. U.S. 30-year carry+roll annualized is 6.0, the highest in the list, but its reliability corr is only 0.077 and win rate 53%, indicating marginal effectiveness. U.S. 10-year carry is 5.35 with corr 0.113 and win rate 56%; U.S. 5-year carry 5.067 and 2-year carry 5.12, with the 2-year corr 0.326 and win rate 68%, showing better static attributes. In other words, on the U.S. curve, if yields stay put, 2-, 5-, 10- and 30-year tenors all offer nontrivial carry, with the 30-year nominally the highest; but if yields continue to rise, the 30-year price sensitivity is greatest and its carry is most likely to be eaten away. Given the current unpredictability of the long end, static yield alone does not justify preferring the long end.

For Canada, one must be more restrained on direction. CA 2Y, 5Y, 10Y and 30Y are all labeled ≈random walk, so direction over the next 3 weeks is unpredictable and no judgment is offered. Although the entire Canadian curve has risen in the past 5, 20 and 60 days and z-scores are near or above 3, the statistical evidence does not support confidently extrapolating this strength into a reliable directional signal. Canadian rates can be described as being at high levels, with a steeper curve and high sensitivity to global inflation and U.S. rates, but this does not justify concluding they will continue to rise.

What can be discussed for Canada is carry and relative shape. CA 30Y carry is 5.06 and CA 10Y carry is 4.56, and both are marked [proven] with win rates of 54% and 56% respectively. This means that in a yield-stable scenario, Canadian long and 10-year tenors have some holding advantage, especially given absolute yields have risen to 4.28% and 3.94% respectively, so coupon cushioning is thicker than before. But again, if the global bond sell-off continues and Canadian yields follow the U.S. up another step, price losses will still outpace carry accumulation.

On cross-country divergence, the U.S. and Canada did not show a clear directional split this time; it was more a case of synchronous pressure with different predictability. The U.S. short and mid curve has validated momentum signals that allow continued upward bias; Canada, despite stronger moves and higher quantiles, statistically remains in an unpredictable zone. For medium-term systems this distinction is critical: both experienced rapid recent yield increases, but continuation evidence is clearer in the U.S. curve than in Canada. From a research perspective, clearer signals over the next 3 weeks remain on the U.S. curve.

On curve shape, both U.S. and Canada maintain positive slopes, with Canada steeper. U.S. 2s10s is +39bp, Canada +63bp. A positive slope implies that holding short-to-mid tenors typically provides better roll-down conditions, particularly in the 2-to-5-year and 5-to-10-year segments. Combining carry+roll data, U.S. 2-, 5- and 10-year static hold metrics are decent, with the 2-year most balanced at carry 5.12, corr 0.326 and win rate 68%. But direction risk cannot be ignored: the current directional judgment for the U.S. 2-year is up, and if yields continue to rise even short-dated bonds will incur losses, though duration is shorter and price impact is relatively more controllable.

Compressing the above into a one-line operational research conclusion: Over the next roughly 3 weeks, the U.S. short and mid curve is biased toward higher yields; the U.S. 30-year and the entire Canadian curve are unpredictable and no directional judgment is offered; looking only at static holding characteristics, the current high-yield environment has made carry attractive again, but carry only works when rates are range-bound or volatility compresses — if yields continue to surge, carry is a buffer, not a reversal signal.

3. Risk Warnings

First, inflation news still dominates the global bond market. Over the past week, from U.S. CPI and PPI to ECB hikes and the ECB President emphasizing a longer-lasting inflation shock, to U.K. inflation risk warnings, the information flow has reinforced a narrative of higher rates staying longer. If this narrative continues to ferment, the upward bias for the U.S. short and mid curve may be reinforced and the long end may remain under pressure, though long-end direction remains unpredictable.

Second, long-end volatility has already widened materially. U.S. 10-year is near 5% and the 30-year at 5.37%, with headlines repeatedly noting bond market gets spooked and selloff shows no sign of diminishing. This indicates long-bond pricing is simultaneously influenced by inflation, supply, term premium and risk appetite. Even if carry is nominally higher, long-end price swings can more easily erase holding gains.

Third, although the Canadian curve rose in sync with the U.S., statistics do not support simple directional extrapolation. Interpreting recent strength as simply continuing risks exceeding the evidence boundary. Canada is better viewed as high yields, a steeper curve and some carry cushion, not as a deterministic directional call.

Fourth, credit spreads are relatively calm now; HY and IG 20-day changes are small, but this does not mean duration risk has been fully absorbed. If rates continue to rise rapidly, credit assets could shift from being mainly rate-driven to a co-movement of rates and spreads.

This report is for research purposes and not investment advice.